A Motion Control and Obstacle Avoidance Algorithm for Redundant Manipulators

碩士 === 國立中山大學 === 海下技術研究所 === 86 ===   This thesis addresses the problems of motion control and obstacle avoidance for redundant manipulators. This type of manipulators have more degrees of freedom than that of the workspace Therefore, we can not use the inverse of manipulator's Jacobian to co...

Full description

Bibliographic Details
Main Authors: Chiu, Guay-Ming, 邱貴明
Other Authors: Wang, Chau-Chang
Format: Others
Language:zh-TW
Published: 1998
Online Access:http://ndltd.ncl.edu.tw/handle/33619897387285855253
Description
Summary:碩士 === 國立中山大學 === 海下技術研究所 === 86 ===   This thesis addresses the problems of motion control and obstacle avoidance for redundant manipulators. This type of manipulators have more degrees of freedom than that of the workspace Therefore, we can not use the inverse of manipulator's Jacobian to command the end-effector. However, the extra degrees of freedom can be used to carry out obstacle avoidance or other secondary tasks besides the primary one.   In this thesis, we will develop an algorithm for motion control and obstacle avoiance by adopting jont anglestaking from a geometrically similar articauated elastic structure.When the tip of this structure is stretched to a desired position. each joint in the structure will deform in such a way that the potential energey stored in the system is minimized. We can command the joints of the manipulator move in the same fashion to reach the desired position. Because the only assumption made in this approach is that each of the forces involved are conservative The same framework can be used to incorporate other forms of conservative force, such as static electric force. In order to carry out obstacle avoidance, we assign electric charges of the same polarity to the links and the obstacles. From physics we know that two charges of the same polarity can not approach each other infinitely clse. Therefore, the artificial electric forces introduced can guarantee a collision-free motion of the manipulator.   Though only planar manipulators and obstacles are studied in this thesis, we still develop a 3D interface under X-Window on PC Linux workstation to expedite the numerical simulations. A 16-DOF planar manipulator operating in a 4-wall corridor environment is creared as a testbed. Results of the numerical simulations indicate that the proposed control algorithm is capable of carrying out motion control and obstacle avoidance in real-time on a Pentium 133 PC workstation.